Subaru Weak Lensing Shear and Magnification Analysis of CLASH Galaxy Clusters

Keiichi Umetsu (ASIAA, Taiwan) with the CLASH team (arXiv:1404.1375)

Cluster Lensing And Supernova survey with Hubble

CLASH Objectives & Motivation

Before CLASH (2010), deep-multicolor Strong (HST) + Weak (Subaru) lensing data only available for a handful of strong-lens clusters

Umetsu+11a, Postman+12

c2D= 6.2 ± 0.3

Umetsu+11b

<c3D> ~ 3

60% superlens bias

Total mass profile shape: consistent w self-similar NFW (cf. Newman+13; Okabe+13) Degree of concentration: maximum superlens correction not enough if <cLCDM>~3?

s

cc r

rc 200200 :=

A 524-orbit HST Treasury Program to observe 25 clusters in 16 filters (0.23-1.6 μm) (Postman+CLASH 12)

CLASH: Observational + Theory Efforts

Wide-field Subaru imaging (0.4 - 0.9 μm) plays a unique role in complementing deep HST imaging of cluster cores (Umetsu+14, ApJ, arXiv:1404.1375)

MUSIC-2 (hydro + N-body re-simulation) provides an accurate characterization of CLASH sample with testable predictions (Meneghetti+14, arXiv:1404.1384)

HST survey complete

CLASH X-ray-selected subsample (0.18<z<0.9)

• X-ray morphology + Tx selection – Tx > 5keV– Small BCG to X-ray-peak offset, σoff ~ 10kpc/h– Smooth regular X-ray morphology

• CLASH theoretical predictions (Meneghetti+CLASH 14)– Composite relaxed (70%) and unrelaxed (30%) clusters– Mean <c200c>=3.9, σ(c200c) = 0.6, c200c=[3, 6]– >90% of CLASH clusters to have strong-lensing features

Optimized for radial-profile analysis (R>2 σoff ~ 20kpc/h)

Lensing Shear and Magnification

• Shear Geometric shape dist: δe+ ~ γ+

• Magnification Flux amplification: µF Geometric area dist: µ∆Ω

Sensitive to “total” matter density

Un-lensed Lensed

)(;21 crit RΣ=Σ+≈ κκµ

)()()(crit RRR Σ−<Σ≡∆Σ=Σ +γSensitive to “modulated” matter density

Shear doesn’t see mass sheet Averaged lensing profiles in/around LCDM halos (Oguri+Hamana 11)

crit/)( ΣΣ= Rκ crit/)( Σ∆Σ=+ Rγ

• Tangential shear is a powerful probe of 1-halo term, or internal halo structure.

• Shear alone cannot recover absolute mass, known as mass-sheet degeneracy

Total Modulated

CLASH-WL Results (1)

Ensemble-averaged internal halo structure:• Halo mass density profile, <∆Σ(R)>• Degree of mass concentration, <c200>from stacked WL-shear-only analysis of CLASH X-ray-selected sample (16 clusters)

Umetsu+CLASH 14, ApJ, accepted (arXiv:1404.1375)

Stacked halo density profile ∆Σ(R)Stacked shear-only analysis provides a net 1-halo-only constraint (γ+,2h<1e-3)

NFW an excellent fit (PTE = 0.66)• M200c = (1.3+/-0.1) 1015Msun• c200c = 4.01 (+0.35, -0.32)at <z>=0.35• bh= 9 +/- 2 (WMAP7+Tinker10)

Corresponding to θEin=(15” +/- 4”) at zs=2, consistent w SL analysis, <θEin>~20” (Zitrin+14, in prep)

Consistent w a family of density profiles for collisionless DM halos (NFW, variants of NFW, Einasto)

Mis-centering effect on γ+negligible at >10σoff~100kpc/h

Integrated constraints on c(M200,z)

M14 (MUSIC-2): σ8=0.82 Bhat13: σ8=0.8Duffy08: σ8=0.8DeBoni13: σ8=0.78

Variance in theory due to different cosmology and mass resolution

35.001.4 35.0

32.0200

=

= +−

zc

• Excellent agreement w CLASH predictions (M14), c200c~4

• Consistent w Bhatt13, Duffy08 (relaxed) predictions at 1σ, c200c~3.6

Umetsu+CLASH14

Theoretical predictions for stacked c(M,z)

CLASH-WL vs. c-M relations

35.001.4 35.0

32.0200

=

= +−

zc

Umetsu+CLASH14

At low M200c, X-ray selection picks up clusters with higher concentrations (Meneghetti+14)

M14 (MUSIC-2): σ8=0.82 Bhat13: σ8=0.8Duffy08: σ8=0.8DeBoni13: σ8=0.78

CLASH-WL Results (2)

• Individual mass profile reconstruction of 20 CLASH clusters from joint Shear+Magnificationanalysis (16 X-ray + 4 lensing clusters)

Umetsu et al. 2014, ApJ, accepted (arXiv:1404.1375)

Method: Combining Shear & MagnificationBayesian joint-likelihood approach (Umetsu+11a; Umetsu 13)

Non-parametric Σ(R) solutionShear + magnification

• Mass-sheet degeneracy broken• Total statistical precision improved by ~20-30%• Calibration uncertainties marginalized over:

A1689 (z=0.18)

CLASH-WL: Joint Shear + Magnification Analysis

M200c=6e14Msun/h (z=0.19) M200c=20e14Msun/h (z=0.45)CLASH low mass CLASH high mass

Shear-magnification consistency: <χ2/dof> = 0.92 for 20 CLASH clusters

Mass Density Profile Dataset

Umetsu et al. 2014, ApJ, accepted (arXiv:1404.1375)

Shear-Magnification ConsistencyM(<r) de-projected assuming spherical NFW (20 CLASH clusters)

Internal systematic uncertainty in the overall mass calibration, empirically derived to be about +/- 8%

Umetsu+CLASH14, arXiv:1404.1375

Umetsu et al. 14, ApJ, accepted (arXiv:1404.1375)

Comparison with WtG @R=1.5Mpc17 clusters in common (Subaru):• WtG: shear-only (Applegate+14),

NFW c200c=4 prior• CLASH: shear + magnification,

NFW log-uniform: 0.1<c200c<10

Un-weighted geometric mean mass ratio (<Y/X> = 1/<X/Y>)• <MWtG/MCLASH> = 1.10• Median ratio = 1.02

Systematic uncertainty in the overall mass calibration of 8% from shear-magnification consistency (Umetsu+14)

No mass dependent bias

Umetsu+CLASH14

CLASH Comparisons with X-ray massesChandra HSE / Subaru-WL XMM HSE / Subaru-WL

Donahue+CLASH14, ApJ, accepted (arXiv:1405.7876)See also Sereno+14

X-ray to WL comparison at R=0.5Mpc • b = 1 - <MChandra/MWL> = 0.05 +/- 0.07 (11 clusters)• b = 1 - <MXMM/MWL> = 0.16 +/- 0.06 (14 clusters)X-ray to WL comparison at r500 [no aperture correction]• b = 1 - <MChandra/MWL> = 0.09 +/- 0.12 (20 clusters)• b = 1 - <MXMM/MWL> = 0.41 +/- 0.07 (16 clusters)

Un-weighted sample mean:

Suzaku-X HSE vs. Subaru WL

Okabe, Umetsu et al.14, PASJ, in press (arXiv:1406.3451)

500c200c

2500c

virial“WL aperture” used

Independent Suzaku-HSE vs. Subaru-WL results, consistent with XMM-HSE vs. Subaru-WL of CLASH collaboration

Shear + Magnification + Strong Lensing

Total mass density profileΣ(R) = Σ1h(R) + Σ2h(R)

around the CLASH “X-ray-selected” sample

)()()|(

)|( mmhh1

hm rMbMr

Mr ξρ

ρξ +=

Clustering of matter around halos with M:

2h term1h term

Averaged cluster (1h) + LSS (2h) from WL shear + magnification + SL

Strong lensing Weak lensingAdding HST SL tightly constrains the inner density profile (R<100kpc/h)

Inner mass profiles from SL follow 1h prediction from outer WL-shear information

Recovered mass-sheet (LSS), consistent w the shear-based halo model prediction, bh=9 +/- 2 (WMAP7+Tinker10)

Umetsu+2014b, in prep

Comparison with pre-CLASH results

CLASH X-ray-selected sample• M200 = 1.3e15Msun• c200 = 4.0• θE ~ 15” (zs=2)• ν=3.8 (bh~9)

Umetsu11b sample• M200 = 1.7e15Msun• c200 = 6.1• θE ~ 36” (zs=2)• ν=4.1 (bh~11)

• C200 vs θE relation, consistent with triaxial CDM halos (Oguri+12)• Similar ν (MAH), similar Σ in outskirts (Diemer & Kravtsov 14)• Increased Σ at R<0.5Mpc/h, consistent w orientation bias (Gao+12)

Umetsu+14, ApJ, accepted (arXiv:1404.1375)

CLASH-WL Summary

• Ensemble-averaged halo structure ∆Σ (1h) of X-ray-regular CLASH clusters is consistent with a family of standard (collisionless) DM predictions:– M200c= (1.3 +/- 0.1) 1015Msun, <z>=0.35– NFW (PTE=0.66): c200c=4.01 (+0.35, -0.32)– Einasto (PTE=0.51): degree of curvature, αΕ=0.19 +/- 0.07

• The stacked-mean concentration agrees with:– theoretical expectation, <c200c> ~ 3.9, which takes into account

CLASH selection function and projection effects (Meneghetti+14)– measured effective Einstein radius, <θEin>=20” (zs=2), from

independent HST-SL analysis (Zitrin+CLASH 14, in prep)

CLASH-WL Summary (contd.)

• Consistent geometric shear vs. magnification measurements allow for accurate cluster mass profile measurements for 20 CLASH clusters with +/-8% systematic mass-calibration uncertainty.

• Total matter distribution Σ (1h+2h) recovered from full-lensing analysis (SL + shear + magnification) is consistent with shear-based halo model predictions (bh=9+/2 at M200=1.3e15Msun, z=0.35), establishing further consistency in the context of LCDM.

CLASH Products released

http://archive.stsci.edu/prepds/clash/

• Calibrated and co-added images [HST, Subaru]• Object catalogs [HST, Subaru]

Subaru (S-Cam) product release will be completed by September 2014

Supplemental Slides

High-resolution space imaging with Hubble for strong lensing

SUBARU multi-color maging for wide-field weak lensing

26

Latest simulation vs. CLASH

Figure from Diemer& Kravtsov 2014b

MBU+11

Color samples (A370, z=0.38) COSMOS 30-band photo-z

SUBARU shear strength as a function of magnitude

Medezinski, Broadhurst, Umetsu+11 29

Non-local substructure effectA substructure at R ~ rvir of the main halo, modulating )()()( RRR Σ−<Σ=∆Σ

Known ~10% negative bias in mass estimates from tangential-shear fitting, inherent to clusters sitting in substructured field (Rasia+12)

Magnification bias effects

Depletion

Enhancement

n/µ

Geometric area distortion

Flux amplification

Broadhurst, Taylor & Peacock 95

Flux-limited source counts: )()( 11

obs fnfn −− >=> µµ

Cluster masses recovered from lensing analysis

Meneghetti+CLASH 14 i32

Scatter in M2D(R) by halo triaxiality

MUSIC-2 simulation by Massimo

Ensemble-averaged DM halo (1h) density profile

Stacking of weak-lensing signals by weighting individual clusters according to the sensitivity kernel matrix:

with the individual sensitivity matrix

defined with the total covariance matrix

With “trace-approximation”, averaging is interpreted as

Umetsu+CLASH 14, arXiv:1404.1375

20 CLASH clusters in Umetsu+1416 X-ray-selected clusters• 15 clusters from 8.3m

Subaru Telescope• 1 southernmost cluster

(RXJ2248) from 2.2m ESO/MPG

• 0.18 < z < 0.69

4 high-magnification clusters• All 4 clusters from 8.3m

Subaru Telescope

<χ2/dof> = 0.92 for 20 CLASH clusters

Mass Comparisons @ R=1.5Mpc Merten+CLASH 14

• <SaWLenS / WL> = 0.96• <WL / WtG> = 0.91• <SaWLenS / WtG> = 0.88

WL (Umetsu+14) shear+mag (Subaru)SaWLenS (Merten+14) SL + shear (HST+Subaru)WtG (Applegate+14) shear (Subaru)

Un-weighted geometric mean mass ratios (<Y/X>=1/<X/Y>)

Note: WL mass calibration uncertainty of 8 percent